Samer Abboud Signal Processing Final Report 29/5/2014

There are there objectives to this report, as we will see in the three following problems. The first objective is to build a VI to find the magnitude response of each filter while changing the 5 designs of filters(Elliptical, Chebychev, Inverse Chebychev, Butterworth and Bessel) and changing the filter types (Lowpass, Highpass, Bandpass, Bandstop). The second objective which is seen in problem 2 is to generate a pulse with uniform white noise, passing it through a median type filter.

Theoretical Part 1:
In problem 1 we have a case structure for the 4 different types of filters. Those four types ( Lowpass, Highpass, Bandpass, Bandstop) require nearly the same input, changing slightly. the inputs are filter design, filter type, impulse response, high and low frequencies, the order of the filter, sampling rate and the attenuation and ripples in those filters that need those inputs. At the output of the case structure, we are transforming the signal into a Fourier transform and passing it through an array subset, multiplying with a logarithm of base 10, passing it through a bundle to finally be displayed on the waveform graph. All this is put inside a while loop with a 50 ms wait.
Theoretical Part 2:
In problem 2 we are inserting a certain number of samples into a pulse VI as well as a uniform white noise VI, while specifying some characteristics like width and delay and noise level. After we add the pulse to the uniform white noise we apply one side of the signal into a build array and another side into a median filter that also fills into the same build array to be finally seen on a waveform graph
Simulation and Analysis Part 1:
In problem 1 we can change the case structure to see the 5 designs of filters (Lowpass, Highpass, Bandpass, Bandstop) each in a different filter type(Lowpass, Highpass, Bandpass, Bandstop). The output is seen on the waveform graph clearly.
Simulation and Analysis Part 2:
In problem 2 after we add the pulse to the uniform white noise we see how the median filter has filtered the noise on the waveform graph. we have the freedom to change the variables like the samples and the amplitude, width and delay noise level as well as the median filter rank.

In conclusion, we can see that each filter works best with some certain characteristics, in addition some certain signals work best while passed through filters, as well as the type greatly affects the outcome.